ADHD is a major public health problem with severe long term outcomes. Existing treatments do not achieve long term change, which has increased the urgency of discovering biomarkers, biological bases of heterogeneity, and developmental specificity of effects. The proposed project would test focused theories of aberrant developmental organization of brain networks in ADHD in a longitudinal design. It would also seek to identify subtypes of ADHD based on brain networks. Two latest-generation methods will be utilized: Diffusion tensor imaging (DTI) and functional connectivity MRI (fcMRI). Use of these methods is directly in line with the NIH Human Connectome guidelines, so the data set will also serve as a resource in the eventual creation of a child connectome. A total of 490 children age 7-10 would be enrolled prior to data loss to attrition and motion artifact. They will be followed annually in three waves, yielding an age based accelerated longitudinal design spanning ages 7-12 years. 70% of the sample will have ADHD (all subtypes) and 30% will be typically developing comparison youth. Community- based recruitment is emphasized so that some treatment naove scans will be obtained. Clinical change, treatment use, and psychosocial moderators will be tracked. Selected cognitive measures will be used to cross validate clinical findings and determine whether mechanistic suppositions about brain circuits hold at the cognitive level of analysis. The major innovations in this project are: (a) evaluates ADHD as a dynamic entity rather than a static phenomenon, (b) the first longitudinal study of functional connectivity or DTI in ADHD, (c) combines functional and structural connectivity measures in ADHD, (d) introduces sophisticated usage of graph theory in order to determine biological subtypes of ADHD based on brain networks. The major significance is that it will enable new discoveries about brain networks in ADHD that can enable significant advances in describing neurobiological markers of liability (those that persist regardless of symptom change) and markers of course of illness (those that are acutely sensitive to phenotypic change or predict outcome or treatment response). Those data can inform new models of assessment, individualized treatments, and treatment monitoring. If the project is successful it will represent an important step forward in mapping integrative, paradigmatic models of ADHD mechanism using the newest insights about brain development.